FREEWAY RAMP MERGING PROCESS OBSERVED IN CONGESTED TRAFFIC: LAG VEHICLE ACCELERATION MODEL

摘要

The freeway merging manoeuvres represent a complex process and involve lane changing, acceleration, deceleration, and merging into a gap. This work presents a methodology for collecting field data, analysing freeway merging behaviour data and developing and calibrating a freeway lag vehicle (approaching the ramp area from the freeway) acceleration models. Wide ranges of data were collected using videotape and image processing techniques. Comprehensive traffic surveys were conducted at two entrance ramps in the Tokyo Metropolitan Expressway. A theoretical framework for modeling the freeway lag driver acceleration-deceleration behaviour is presented. It is built on the content of the earlier work in which a general methodology of modeling ramp vehicle acceleration was introduced. This methodology uses the stimuli-response psychophysical concept as a fundamental rule, and is formulated as a modified form of the conventional car-following models. Data collected at the two merging points of the Tokyo Metropolitan Expressway are used to calibrate the hypothesized freeway lag vehicle acceleration models. It is found that the surrounding freeway leader vehicle and especially ramp vehicle affect significantly the freeway lag vehicle acceleration behaviour. Furthermore, the average speed of merging vehicles (freeway leader, ramp and freeway lag vehicle) of a simulation program, which has been developed using the outcome of the developed acceleration model, shows reasonable consistency with the collected data from instrumented vehicle. Finally, it is believed that the developed acceleration model of lag vehicles, in addition to the ramp driver acceleration model proposed in earlier work, constitutes a complete picture of driver interaction behaviours at the time of merging under congested traffic conditions.